
FPGA Acceleration of Convolutional Neural Networks (CNNs)
White Paper FPGA Acceleration of Convolutional Neural Networks Overview Convolutional Neural Networks (CNNs) have been shown to be extremely effective at complex image recognition problems.
WaveBox contains RF modules that are built around the powerful AMD RFSoC devices. We bring analog signals in, but it’s the direct RF chip approach that packs so much analog and digital functionality that many racks of previous equipment are no longer necessary!
With WaveBox, you’ve got options in the number of modules and type. You may want a box with just one RFX module to begin development, then add more. Or, you may need RF + FPGA, with the RFX module (or two) handling ADC and moving digitized packets over to an XUP-VV8 FPGA card for further digital processing.
Processing analog RF used to require large, bulky equipment. For many applications, from Satcom to Quantum computing, much of this equipment has been reduced to a single chip, the revolutionary AMD RFSoC. However, there remains equipment like downconverters, amplifiers, and filtering that can take up space and add complexity.
That’s where we have brought in WaveBox, a modular RF processor with built-in mixing and amplifier/filtering in single chassis. The RFX modules that make up WaveBox are built from a COTS perspective—taking advantage of higher volume standard modules, avoiding ruggedized form factors, and yet retaining high performance in the critical analog signal path.
Our focus is on satellite communications ground stations and test/measurement including quantum computing. Any deployment that can take a rackmount chassis is an excellent fit for the compact WaveBox form factor.
High analog port density, with an option for RFX 881/771 to have even more ports dedicated to ADCs or DACs
Configure modules to use your own filtering, or choose our built-in 1st, 2nd, or 3rd Nyquist amplifier + filter
Up to 35 GHz to/from your antenna, converted to multiple channels as your application requires it.
Whether you need CPU-based or standalone architectures, WaveBox gives you a digital bandwidth advantage
Getting data formatted, encrypted, and recorded to disk can involve significant development time using specialist firmware engineers. Reduce your time to market by using our partners for these components. They are familiar with the RFX family and are ready to take on your unique requirements!
This 1U enclosure is designed for users who don’t need a host CPU connected to the RFSoC modules.
Key features:
This 2U enclosure is designed with users who need CPU host support, connected to the RFX through PCIe.
Key features:
Our technical sales team is ready to provide availability and configuration information, or answer your technical questions.
"*" indicates required fields
White Paper FPGA Acceleration of Convolutional Neural Networks Overview Convolutional Neural Networks (CNNs) have been shown to be extremely effective at complex image recognition problems.
Panel Discussion How Today’s FPGAs are Taming the Data Deluge Problem From Gen5 to AI, NOCs to RF at the Edge Watch the recording for
BittWare Webinar High Performance Computing with Next-Generation Intel® Agilex™ FPGAs Featuring an Example Application from Barcelona Supercomputing Center Now Available On Demand (Included is recorded
PROVA-C Appliance Running GENEM-C Software PROVA-C Appliance Runs GENEM-C 100G Network Test Application 4× 100G 1U Appliance Built on BittWare Built on BittWare The PROVA-C